How CS2 Baccarat Actually Works
Baccarat is the simplest major card game in the casino, and that simplicity is the whole point. You place one of three bets — Banker, Player, or Tie — and then you do nothing. Two hands are dealt two cards each. Totals use only the last digit, so a 9 and a 7 total 16, which counts as 6. Whichever hand finishes closest to 9 wins.
The part that matters mathematically is the third-card rule. Whether either hand draws a third card is determined by a fixed, published table — not by any choice. The Player hand resolves first: it draws on totals 0–5 and stands on 6–7. Only then does the Banker hand act, and critically, the Banker's drawing decision depends on what the Player drew. That asymmetry is the entire source of Banker's advantage, and it's why the two bets don't have identical odds despite looking symmetrical.
The key structural fact: the Banker hand acts second and reacts to the Player's third card. Acting on more information wins more often — that's true in poker, and it's true here. Baccarat just automates it.
The Three Bets and Their Real Numbers
Here are the actual 8-deck probabilities, which are standardized across virtually every implementation because the drawing rules are fixed:
| Outcome | Probability (all hands) | Excluding ties | Typical payout | House edge |
|---|---|---|---|---|
| Banker wins | 45.86% | 50.68% | 0.95:1 (after 5% commission) | 1.06% |
| Player wins | 44.62% | 49.32% | 1:1 | 1.24% |
| Tie | 9.52% | — | 8:1 | 14.36% |
To put the Banker bet's 1.06% in perspective against the rest of the games covered on this blog:
Same game, three bets — and a 13.5× spread between the best and worst of them.
The Commission Paradox
Here's the thing that trips up almost everyone who looks at baccarat odds for the first time. The Banker bet wins more than half of all decided hands. Not 49-point-something. Not "about half." A genuine 50.68%, every session, forever, because it's structural rather than random.
Now run the math on what that would mean at a clean 1:1 payout. Bet $100 a hand, 1,000 hands, ties excluded:
| Scenario | Wins | Losses | Net result on $100 flat bets |
|---|---|---|---|
| Banker at 1:1 (no commission) | 506.8 | 493.2 | +$1,360 to the player |
| Banker at 0.95:1 (5% commission) | 506.8 | 493.2 | −$1,174 to the player |
That's the entire story of the commission in one table. Without it, Banker is a player-positive bet — the casino would be running a game it loses at scale. The 5% cut on winning Banker bets is not an arbitrary fee or a convenience charge; it is the specific, calibrated mechanism that converts a +1.36% player advantage into a −1.06% house edge.
Why this matters practically: a lot of "baccarat systems" sold online are built on the observation that Banker wins more often. That observation is completely correct — and completely useless, because the commission is priced to exactly that number. Any system whose premise is "bet Banker because it wins more" is describing a fact the casino already charged you for.
The "No Commission" Trap
This is where a genuinely useful, counterintuitive piece of knowledge lives — and it's one of the few places in baccarat where your table choice actually changes your expected value.
"No commission" baccarat, also marketed as Super 6, does exactly what it says: no 5% fee on Banker wins. It sounds strictly better. It isn't. The rule that pays for it is that when the Banker wins with a total of 6, it pays only 1:2 — half your stake instead of even money.
That single carve-out is worth more to the house than the commission was:
| Table type | Banker house edge | Player house edge | Which bet is better? |
|---|---|---|---|
| Standard (5% commission) | 1.06% | 1.24% | Banker |
| EZ Baccarat (3-card Banker 7 pushes) | ~1.02% | 1.24% | Banker |
| No commission / Super 6 | ~1.46% | 1.24% | Player ← reversed |
✅ The one piece of table-selection advice worth remembering: on a standard commission table, bet Banker. On a no-commission / Super 6 table, bet Player. It's the only situation in baccarat where the correct answer flips, and most guides never mention it because they treat "no commission" as a straightforward upgrade.
Several CS2 and crypto platforms — 500Casino among them — offer no-commission alongside squeeze and speed live formats. If you're picking a table there, this distinction is worth the ten seconds it takes to check the rules panel.
Why the Tie Bet Is the Worst Wager on the Table
Ties happen 9.52% of the time — roughly one hand in every ten and a half. That frequency feels like it should support the typical 8:1 payout. It doesn't, and the gap is where the 14.36% edge comes from.
Work it backwards. For a bet occurring 9.52% of the time to break even, it needs to pay about 9.5:1. At 8:1, you're being paid roughly 84% of fair odds on an outcome that already only shows up once every ten-and-a-half hands.
Expected loss of $14.36 per $100 wagered, versus $1.06 on Banker. You would need to place roughly 13.5 Banker bets to burn the same expected value as a single Tie bet of equal size. The 8:1 payout is doing all the persuasive work here — the underlying frequency doesn't come close to supporting it.
Some platforms offer 9:1 on Tie instead of 8:1, which improves the edge substantially — but even at 9:1 it remains meaningfully worse than either main bet. There is no configuration of standard Tie odds that makes it competitive with Banker.
Why Card Counting Doesn't Work Here
This question comes up constantly because baccarat and blackjack look superficially similar — cards, shoe, dealer, low house edge. But counting works in blackjack and fails in baccarat, and the reason is specific and measurable.
In blackjack, removing a single card from the shoe shifts your expected value by 0.5% to 0.67%. That's an enormous swing relative to a house edge of around 0.5%, which is exactly why counting can flip the game player-positive. Every high card that leaves the shoe genuinely changes what you should do next.
In baccarat, removing any single card shifts the outcome probabilities by less than 0.01%. The reason traces back to the third-card rule again: because no player decisions exist and the drawing rules are rigidly fixed, card composition barely influences the result distribution. There's nothing to adapt, because you never make a choice that composition could inform.
| Metric | Blackjack | Baccarat |
|---|---|---|
| EV shift per removed card | 0.5% – 0.67% | under 0.01% |
| Gain from perfect counting | Can flip the edge positive | ~0.01% per hand |
| Player decisions available | Hit, stand, double, split | None |
An ideal counter — playing perfectly, betting only when the count is favorable, never making an error — gains roughly 0.01% per hand. That is one basis point. Not one percent, not a tenth of a percent. It is so far below the 1.06% house edge that no realistic bankroll, session length, or bet spread turns it into an advantage.
On provably fair CS2 and crypto baccarat the point is moot anyway: the shoe is regenerated from a fresh server seed, so there is no persistent physical deck composition to track in the first place.
A note on edge sorting: the famous Phil Ivey case (roughly $20 million, later reclaimed after courts sided with the casinos) is often cited as proof that baccarat is beatable. It isn't card counting — it's exploiting physical printing defects on the card backs to identify values before the deal. It requires a defective product, physical cards, and cooperation from the house on shuffle procedure. None of those exist in an online provably fair implementation.
Baccarat vs Blackjack — Decisions vs None
Both games sit at the low-house-edge end of the casino, so the choice between them comes down to whether you want your input to matter:
| Aspect | Baccarat | Blackjack |
|---|---|---|
| Player decisions | None after the bet | Hit, stand, double, split |
| Skill impact on edge | Zero | Measurable — basic strategy lowers it to ~0.5% |
| House edge | 1.06% (Banker) to 14.36% (Tie) | ~0.5% with correct play, far worse without |
| Learning curve | None — pick a bet | Real — basic strategy chart required |
| Round pace | Very fast, no decision delay | Slower, each decision costs time |
The honest framing: blackjack rewards you for learning something, baccarat doesn't punish you for not learning anything. If you play blackjack without basic strategy, your effective edge is considerably worse than baccarat's Banker bet. If you're not going to memorize the chart, Banker in baccarat is genuinely the stronger default.
What Doesn't Work
Scorecards and streak tracking
Land-based baccarat tables famously hand out scorecards for tracking Banker/Player results, and online implementations often replicate the "roads" display. It's a genuine tradition — and it's decorative. Each hand is statistically independent. On a provably fair CS2 table the shoe is regenerated from a new seed, so previous results contain literally zero information about the next hand.
Martingale and progression systems
Doubling after losses feels like it should exploit Banker's 50.68% win rate. It doesn't change expected value by a single cent — it only reshapes the distribution of outcomes, front-loading small frequent wins against a rare catastrophic losing streak. The same conclusion applies here as it does in Crash and Dice: progressions change variance, never the edge.
"Betting the streak" vs "betting against the streak"
Both camps exist, both are equally wrong, and they can't both be right — which should be the tell. Since hands are independent, a run of six Bankers has no bearing on the seventh hand whatsoever. Banker is the better bet on hand seven for exactly the same reason it was on hand one: the fixed drawing rules, not the history.
Mixing in Tie bets "for value"
A common pattern is playing Banker consistently but sprinkling in occasional Tie bets for the big payout. Every Tie bet you place drags your session's blended house edge sharply upward. Ten Banker bets at 1.06% plus one Tie bet at 14.36% gives you a blended edge of roughly 2.27% — you've more than doubled your cost with a single hand.
🎁 Bet Banker, skip Tie, check the commission rule — use code BINROLL
That's the whole strategy, and it's genuinely all there is. Code BINROLL gives bonuses on all 18 tested baccarat sites.
FAQ
⚠️ Gamble Responsibly
A 1.06% house edge is low, but it is still an edge — baccarat carries guaranteed negative expected value over enough hands, and its fast round pace means you can wager a surprising amount in a short session. The low edge is a reason to prefer Banker over Tie, not a reason to play longer or bet bigger. Set a session budget before your first hand. Visit BeGambleAware for free support. 18+ only.